French localization (csl 1.0)

Gracile

Hello,
I translated csl 1.0 locales into French (http://bitbucket.org/gracile/zotero-fr-locales/downloads). EDIT: http://bitbucket.org/gracile/csl-locales-fr-fr/src

1) I have a problem with ordinals. In French (for a similar problem in Spanish, see this thread) "first" can be: "premier" (masculine), "première" (feminine), "premiers" (masculine and plural), "premières" (feminine and plural).
E.g.: "1st ed." is "1^re éd." but "January 1st" is "1^er janvier"

There has been some debates on csl implementation of this but it has been - and I fully understand that - considered as very difficult to implement...

2) I'm not sure I've fully understood the "open-quote"/"close-quote" and "open-inner-quote"/"close-inner-quote" behaviour. (in spite of this wonderful document) I'll test that.

Rintze

1) this indeed isn't covered by CSL 1.0.

2) do these examples help?
http://www.wikipublisher.org/wiki/Wikipublisher/TypographicQuotes

Gracile

1) Just want to keep a writing note of this issue (shared with other languages and without any easy workaround - except a full manual review, (hum... thinking while writing) search and replace might be useful)

2)I had seen these examples. I think the locales-fr-FR.xml above is correct. The problem is with flip-flopping actually (uh, I should have said that in my first post..)

Is the locale okay for you?

Rintze

Flipflopping of quotes ensures that you get properly nested output like
“Speak ‘friend’, and enter” instead of
“Speak “friend”, and enter” or
‘Speak ‘friend’, and enter’ or
‘Speak “friend”, and enter’, regardless of how quotes are provided in the input data.

I'll review your translations and commit them to the CSL locale repository if they look okay.

Gracile

Okay. Thanks for the explanation: that was not as clear in the upgrade notes.

Gracile

One more question: is it possible to add superscripted characters in the localized terms?
The short form of issue is "n°" but it should be "n^o". (Source)
The degree symbol is used because it is more practical (direct access with keyboard) but that's not orthodox.

That would be great because the pluralized short form is n^os although there is no such pluralized short form for issue in csl (like for pages where p. and pp. are possible)

Rintze

Also, just thinking aloud about issue (1), we might be able to cover this case transparently by a) allowing gender information to be included on nouns, and b) allowing gender-specific terms to be defined. For example, locales-fr-FR.xml could include:

<term name="edition" gender="feminine">
<single>édition</single>
<multiple>éditions</multiple>
</term>
<term name="month-01" gender="masculine">janvier</term>

<term name="ordinal-01" gender="feminine">re</term>
<term name="ordinal-01" gender="masculine">er</term>
<term name="ordinal-02">e</term>
<term name="ordinal-03">e</term>
<term name="ordinal-04">e</term>
I'm not sure though if this covers all cases.

Gracile

That's brilliant! I was thinking about b) but not a) which is a smart way to address this issue.

Rintze

I'm not sure if citeproc-js can handle it, but did you try escaping the superscript HTML tags? E.g.:
<term name="issue" form="short">n& lt;sup& gt;o& lt;/sup& gt;</term>
(you have to remove the four spaces between "&" and "lt;" or "gt;" I added to prevent the forum software to parse the HTML escape characters into HTML, see also http://www.theukwebdesigncompany.com/articles/entity-escape-characters.php )

MHSmith

Funny, I have also been revising the same French translation today. Some solutions are similar, which is reassuring, and others are different, partly because I am following a journal-specific norm.

One practical problem is, I have just realised I don't know what to do with the xml file once it is done: how do I use it? I also tried inserting the code into a citation-style I am working on for the same journal, but Firefox seems to consider any French diacritics (at least in this part of the code) as bad code, and won't install my style with the additional locale stuff. Would be grateful for any suggestions. (I have only been having a go at this for a couple of days, and with scarce xml experience.)

Rintze

Is the locale okay for you?

It has been committed: http://bitbucket.org/bdarcus/csl-locales/changeset/534102a6b8d9

The only thing I changed was to add a period to <term name="ad">apr. J.-C</term>

ajlyon

If the superscript doesn't work, and perhaps even if it does, I would suggest trying a Unicode superscripted O (U+1D52, ᵒ) for the No. term.

MHSmith

Suggestions for more changes:

accédé > consulté (accéder is intransitive)

circa: is this only used for dates? if it is, then "vers" / "v." is the traditional form

s.d. > s. d. — at least if it could be a non-breaking space, but maybe not? Same for s.v. and s.vv.

présenté au > what about feminine and plural references (same problem as premier/premières/premières…)? So "présenté à", although not perfect, would be less problematic, wouldn't it?

plurals:
- réf., vol. (you can't just add an s to an abbreviation in French)
- v. both singular and plural (same as p. and others)

retrieved: not sure what is is used for, but "retrouvé" doesn't look right (would mean something like lost & found)

deuxième: French has both "second" (of two) and "deuxième" (of more), but I suppose that is unfeasible. If so, "deuxième" is OK.

lines and verses (tricky), in English verse applies to biblical refs, and lines both to poetry and prose, right? French has "verset/v." (Bible), "vers/v." (poetry), and "ligne/l." (prose). Any possibility in future CSL developments?

éditeur et traducteur, éd. et trad. (et, not &) is more usual, at least in the humanities.

edité & traduit par > édité (accent) et traduit par

(recipient: of letter?) pour > à

éd. & trad. par > éd et trad. (w/o par)

jan. > janv.

juill. > juil. (an exception to standard abbreviation principles, in order to keep all months at 4 letters max)

PS
Another tricky and pretty illogical French usage is about volumes:
- 3 vol. = trois volumes
- t. 3 or even better t. III = tome trois
(where I work, we use t. 3 for periodicals, and t. III for books)
- 2 t. en 3 vol. (2 vols in 3, two intellectual volumes printed/bound as three physical volumes)

I don't suppose CSL can do anything about converting between arabic and roman numerals, though it would be nice.

Gracile

@Rintze: That's right (apr. J.-C.). I'll try the superscript later.

@MHSmith:

Funny, I have also been revising the same French translation today. Some solutions are similar, which is reassuring, and others are different, partly because I am following a journal-specific norm.

Indeed, I've tried to follow French orthography & typography rules (when they exist: tradition is not always logic... and illogical cases are difficult to implement in csl). But there are improvements to do! Thanks for your suggestions.

accédé > consulté (accéder is intransitive)

Done.

circa: is this only used for dates? if it is, then "vers" / "v." is the traditional form

Not easy. Circa is not used in French journals but very common in library catalogs... "vers"/"v." is right but "environ"/"env". is used too...

s.d. > s. d. — at least if it could be a non-breaking space, but maybe not? Same for s.v. and s.vv.

Done for "s. d." Not sure for "s.v."/"s.vv."

présenté au > what about feminine and plural references (same problem as premier/premières/premières…)? So "présenté à", although not perfect, would be less problematic, wouldn't it?

Done. I agree with your comment on feminine and plural refs but, in this case, it cannot be implemented (gender agreement depends on the variable in zotero...)

plurals:
- réf., vol. (you can't just add an s to an abbreviation in French)
- v. both singular and plural (same as p. and others)

Right. Done. Note that there is a potential conflict here: if "v." is " vers" (poetry), can it be "vers" (circa)? That seems to me conceptually strange, but they are homonyms.

deuxième: French has both "second" (of two) and "deuxième" (of more), but I suppose that is unfeasible. If so, "deuxième" is OK.

You got it. It's too hard to implement.

lines and verses (tricky), in English verse applies to biblical refs, and lines both to poetry and prose, right? French has "verset/v." (Bible), "vers/v." (poetry), and "ligne/l." (prose). Any possibility in future CSL developments?

You're right.

éditeur et traducteur, éd. et trad. (et, not &) is more usual, at least in the humanities.

Done.

edité & traduit par > édité (accent) et traduit par

Done.

(recipient: of letter?) pour > à

Right. Done.

éd. & trad. par > éd et trad. (w/o par)

Indeed. Done.

jan. > janv.

Right. Strange, I'm sure I revised months very carefully. Done.

juill. > juil. (an exception to standard abbreviation principles, in order to keep all months at 4 letters max)

I disagree. A (common) mistake for me (See)

I don't suppose CSL can do anything about converting between arabic and roman numerals, though it would be nice.

It would be a sort of number flipflopping. I don't know if it is possible.

Gracile

I'm not a regular user of bitbucket: I had some difficulties to upload the new xml file. Thus, I've put the updated French locales here: https://bitbucket.org/gracile/locales-fr-fr/downloads

Edit: http://bitbucket.org/gracile/csl-locales-fr-fr/src

MHSmith

Gracile, thanks for changes and answers.

Indeed different authorities do disagree, so I won't argue for the old-fashioned juil. (under the influence of Latin "Jul.") vs juill. The same could be said about "apr. J.-C.", which was traditionally "ap. J.-C.", for the sake of symmetry with "av. J.-C."

True, c.=v.=env.

Also true, in both long and short forms, "vers/v." (circa) is the same as "vers/v." (line), but it can hardly cause ambiguity in any context.

About spaces within abbreviations: they should either be used or not, I think. In French they normally are, so there is no reason why "s. v." and "s. d." shouldn't be treated the same, is there? I am just worried about line breaks: better no space than a breaking space.

Oh, and it seems the first change didn't register, the file still has "accédé".

MHSmith

PS Could someone please tell me how/where the file should be installed? Thanks again.

Gracile

Ok for "vers"/"v." instead of circa.
I'll put non-breaking space for "s. v." too.
Fixed "accessed" as "consulté".
But bitbucket does not "want" to update the file, I don't understand: I'll do that this evening.

PS Could someone please tell me how/where the file should be installed? Thanks again

While Firefox closed:
Unzip zotero.jar which should be (on Windows) in YourFirefoxProfile\extensions\zotero@chnm.gmu.edu\chrome\
Go to: \content\zotero\locale\csl\
You'll need to replace the locales-fr-FR.xml. Then rezip (with .jar extension)

Rintze

Also, you need to be running the Zotero 2.1 beta in order to be able to use a CSL 1.0 locale file.

@Gracile: you don't really need it for now, but bitbucket.org has been designed to be used along with the Mercurial versioning control system (the main benefit would be that it is easy to track and revert changes in your translation file). If you're running Windows, TortoiseHg (http://tortoisehg.bitbucket.org/) is a very nice client.

MHSmith

Thanks to both!

MHSmith

Oh no, 2.1b doesn't work with the Zotero test tool (chrome://zotero/content/tools/csledit.xul). Pity.

And how on earth do I make a .jar? Can I just zip up and change the extension?

dstillman

csledit.xul works in 2.1b2, now available.

Yes, a .jar is just a .zip file with a different extension (and no compression, ideally).

http://www.zotero.org/support/dev/modifying_zotero_files

MHSmith

Great, thanks! Funny, I downloaded Zotero 2.1 today, and got b1, not b2. Wonder why It isn't posted on the homepage.

dstillman

Beta 2 was released five minutes ago.

MHSmith

Ah, lucky me. But now my style doesn't work anymore, either in clsedit.xul or in MS Word (nor does the one I based it on, "EHESS temporaire").

Gracile

@MHSmith: did your style work with 2.1b1? Zotero 2.1 implements csl 1.0: zotero 2.0 styles do not work with zotero 2.1. See: http://citationstyles.org/downloads/upgrade-notes.html#updating-csl-0-8-styles

MHSmith

Not quite sure whether it worked in b1, since I moved on to b2 almost immediately. But I believe it is csl 1.0 (at least the file it is based on says so). Is there any documentation on the differences between the two?

BTW, out of four Chicago styles, one (Note w/o bibliography) also seems to be broken. — And so does MHRA Note w/o bibliography.
But surely this isn't the right thread to carry on this discussion, is it?

Gracile

If you're not sure, I think it's a csl 0.8 style.
Docs:
http://citationstyles.org/downloads/primer.html
http://citationstyles.org/downloads/upgrade-notes.html
http://citationstyles.org/downloads/specification.html

[Indeed, I think you should start a new thread to keep this one focused on French localization]

MHSmith

From a few elements it really does seem I have 1.0, but I'll check for any mistakes. Thanks again, and maybe more later.

Gracile

Did you try to validate your style: http://citationstyles.org/downloads/primer.html#validation

MHSmith

I am afraid not, thanks for the suggestion. Now it's done, I have 46 errors. So better off to bed and back to work tomorrow.